一种幅度信息辅助多伯努利滤波算法

发布时间：2018-11-13 09:13

【摘要】：在许多多目标跟踪场景中,目标返回的幅度通常强于虚警杂波返回的幅度。通过建立更加准确的包含幅度信息的目标和虚警杂波似然函数,可提高多目标估计精度。该文提出一种基于随机有限集的幅度信息辅助多伯努利滤波(Amplitude Information Assistant Multi-Bernoulli Filter,AIA-MBer F)算法。该算法通过建立幅度似然函数将幅度信息引入到多伯努利滤波的更新过程中,并给出针对线性和非线性模型的高斯混合(Gaussian Mixture,GM)和序贯蒙特卡洛(Sequential Monte Carlo,SMC)实现方法。仿真结果表明,该滤波算法相比于传统多伯努利滤波(Multi-Bernoulli Filter,MBer F)无论GM还是SMC实现都可获得更加准确稳定的目标数和对应的目标状态估计。
[Abstract]:In many multi-target tracking scenarios, the return amplitude of target is usually stronger than that of false alarm clutter. The accuracy of multi-target estimation can be improved by establishing a more accurate likelihood function of target and false alarm clutter. In this paper, an amplitude information aided multiple Bernoulli filter (Amplitude Information Assistant Multi-Bernoulli Filter,AIA-MBer F) algorithm based on random finite sets is proposed. In this algorithm, amplitude information is introduced into the updating process of multiple Bernoulli filtering by establishing amplitude likelihood function, and the implementation methods of Gao Si mixture (Gaussian Mixture,GM) and sequential Monte Carlo (Sequential Monte Carlo,SMC (SMC) for linear and nonlinear models are presented. Simulation results show that the proposed algorithm can obtain more accurate and stable target number and corresponding target state estimation than the traditional multi-Bernoulli filter (Multi-Bernoulli Filter,MBer F), regardless of the GM or SMC implementation.
【作者单位】：北京航空航天大学电子信息工程学院;
【基金】：国家自然科学基金(61171122;61201318;61471019;61501011) 中央高校基本科研业务费专项资金(YWF-15-GJSYS-068)~~
【分类号】：TN713

【相似文献】